Pump and Dump Cryptocurrency Detection Using Social Media

Domenico Alfano, Domenico Alfano, Roberto Abbruzzese, Roberto Abbruzzese, Domenico Parente

2023

Abstract

The economic implications behind the fluctuation of cryptocurrencies prices, and, more importantly, the complexity of the variables involved in the process, have made price forecasting a very popular topic among researchers. Especially around detecting Pump & Dump events, where investors try to manipulate cryptocurrency owners to either buy or sell making a profit from them. Over the last decade, research has progressed by proposing new metrics (financial and non financial) capable of influencing and tracking the reasons for price fluctuations. Thanks to the advent of social media, major investment communities can be analysed through social channels to create new metrics. With developments in the field of Natural Language Processing, these social channels are used to extract opinions and mood of expert investors and cryptocurrencies owners. We propose to apply those innovative ways of creating metrics and to demonstrate that, taking these generated metrics into account, can significantly outperform other existing Pump & Dump detection methods. Moreover, to measure how each created metric contributes to the detection, a game theory approach called SHapley Additive exPlanations and a method that explains each prediction using a local, interpretable model to approach any black box machine learning model called Lime will be used.

Download


Paper Citation


in Harvard Style

Alfano D., Abbruzzese R. and Parente D. (2023). Pump and Dump Cryptocurrency Detection Using Social Media. In Proceedings of the 12th International Conference on Data Science, Technology and Applications - Volume 1: DATA; ISBN 978-989-758-664-4, SciTePress, pages 235-240. DOI: 10.5220/0012059300003541


in Bibtex Style

@conference{data23,
author={Domenico Alfano and Roberto Abbruzzese and Domenico Parente},
title={Pump and Dump Cryptocurrency Detection Using Social Media},
booktitle={Proceedings of the 12th International Conference on Data Science, Technology and Applications - Volume 1: DATA},
year={2023},
pages={235-240},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012059300003541},
isbn={978-989-758-664-4},
}


in EndNote Style

TY - CONF

JO - Proceedings of the 12th International Conference on Data Science, Technology and Applications - Volume 1: DATA
TI - Pump and Dump Cryptocurrency Detection Using Social Media
SN - 978-989-758-664-4
AU - Alfano D.
AU - Abbruzzese R.
AU - Parente D.
PY - 2023
SP - 235
EP - 240
DO - 10.5220/0012059300003541
PB - SciTePress